@rijwind/vane

Read and render .vane weather datasets in the browser. PMTiles, but for weather data: one immutable file per model run, hosted on any static host or CDN, read with HTTP range requests. No tile server.

A .vane file is a Zarr v3 sharded store packed into a single file (spec in spec/). This package is the JavaScript reader plus a set of MapLibre render modes:

• ColormapLayer — scalar fields (temperature, precipitation, clouds, gusts)
• ParticlesLayer — GPU wind particles
• ArrowsLayer — instanced direction glyphs colored by speed
• ContoursLayer — isolines with labels (e.g. pressure)
• ValuesLayer — honest grid-point value labels

Install

npm install @rijwind/vane maplibre-gl

maplibre-gl (>= 4) is a peer dependency. The only runtime dependency is zarrita.

Quick start

import { VaneDataset, ColormapLayer, ParticlesLayer } from "@rijwind/vane";

// Resolve a *_latest.json pointer and open the immutable .vane behind it.
const ds = await VaneDataset.openLatest(
  "https://weather.rijwind.com/knmi_harmonie_nl_latest.json",
);

map.addLayer(new ColormapLayer({ id: "temp", dataset: ds, variable: "temperature" }));
map.addLayer(new ParticlesLayer({ id: "wind", dataset: ds }));

Note the split: ColormapLayer, ParticlesLayer and ArrowsLayer are WebGL custom layers (map.addLayer(layer)); ValuesLayer and ContoursLayer wrap MapLibre style layers and go on the map with layer.addTo(map).

Every layer has setTimestep(t) for time animation, and VaneDataset.getPointSeries(variable, lon, lat) returns the full time series at a point for meteograms.

Live demo with open KNMI / DWD / ECMWF data: https://rijwind.com/weather.

Making .vane files

The Python writer and CLI (GRIB → regrid → quantize → Zarr v3 → .vane) live in the same repo, together with a reference ingest pipeline: https://github.com/Rijwind/vane.

License

Apache-2.0 © Rijwind